Does the pattern of educational inequalities in smoking in Western Europe depend on the choice of survey?
International Journal of Public Health , Volume 59 - Issue 4 p. 587- 597
Objectives: Smoking rates vary according to socioeconomic group. We investigated whether patterns of educational inequalities in smoking prevalence differ across three major European surveys. Methods: Data on smoking came from National Health Interview Surveys (NHIS), the European Community Household Panel (ECHP) and the Eurobarometer (EB). We calculated prevalence ratios by education. We controlled for sex, country, data source and age. We used likelihood ratio tests to determine whether inequalities in each country differed between surveys and whether the association of education and smoking across countries was the same in different surveys. Results: Smoking prevalence tended to be lower in the ECHP than in both other surveys, and was highest in the EB. The pattern of inequalities in smoking also differed between surveys. Statistically significant differences between surveys were found mainly in Southern Europe, where EB-based prevalence ratios often deviated from those in the other two surveys. Conclusions: Relative inequalities in smoking prevalence depend on the survey used. Our results suggest that the NHIS and the ECHP are more reliable sources of information on educational inequalities in smoking than the EB.